u dE/dx measurementasai/work/Particle_detector...CERN Summer Student Lectures 2002 Particle...

CERN Summer Student Lectures 2002Particle Detectors Christian Joram V/1

u dE/dx measurement

u Time of flight

u Cherenkov detectors

u Transition radiation detectors

Particle Identification

π

Κ

µ

p

π

Κ

µ

p



Particle identification is an important aspect of high

energy physics experiments.

Some physical quantities are only accessible with

sophisticated particle identification (B-physics, CP

violation, rare exclusive decays).

One wants to discriminate: π/K, K/p, e/π, γ/π0 ….

The applicable methods depend strongly on the

interesting energy domain.

Depending on the physics case either εxx or εxy

has to be optimized:

Efficiency:

Misidentification:

Rejection:

The performance of a detector can be expressed in

terms of the resolving power

xtagxxx NN=ε

ytagx

yxy NN −=ε

xyxxxyR εε=

Q

yxyx

QQD

σ

−=,

εxx

εxy



Particle ID using the specific energy loss dE/dx

( )222

ln1 γβ

β∝

dxdE

βγ0mp =Simultaneous measurement of p and dE/dx defines mass m, hence the particle identity

Average energy loss for e,µ,π,K,p in 80/20 Ar/CH4 (NTP)(J.N. Marx, Physics today, Oct.78)

π/K separation (2σ) requires a dE/dxresolution of < 5%

e

But: Large fluctuations + Landau tails !

Specific energy loss

p

K

πµ

p

K

πµ

p

K

πµ


Improve dE/dx resolution and fight Landau tails

(M. Aderholz, NIM A 118 (1974), 419)

(B. Adeva et al., NIM A 290 (1990) 115)

1 wire 4 wiresL: most likely

energy lossA: average

energy loss

Don’t cut the track into too many slices ! There is an optimum for each total detector length L.

• Chose gas with high specific ionization• Devide detector length L in N gaps of thickness T.• Sample dE/dx N times

• calculate truncated mean, i.e. ignore samples with (e.g. 40%) highest values

• Also pressure increase can improve resolution, but reduced rel. rise due to density effect !

Specific energy loss (backup)


Example ALPEPH TPC

Gas: Ar/CH4 90/10

Nsamples = 338, wire spacing 4 mmdE/dx resolution: 4.5% for Bhabhas, 5% for m.i.p.’s

log scale !

linear scale !



log p [GeV/c]

∆E

(a.

u.)

log p [GeV/c]

∆E

(a.

u.)

dE/dx can also be used inSilicon detectors

Example DELPHI microvertex detector (3 x 300 µm Silicon)



∆t for L = 1 m path length

σt = 300 psπ/K separation up to 1 GeV/c

Time of flight

cL

tβ

=

12

22−=

L

tcpm

start stop

Combine TOF with momentum measurement

++=

LdL

tdt

pdp

mdm 2γMass resolution

TOF difference of 2 particles at a given momentum

( ) ( )22

212

2222

2221

21 2/1/1

11mm

pLc

pcmpcmcL

cL

t −≈+−+=

−=∆

ββ

Particle ID using Time Of Flight (TOF)

)( 0βγmp =


Example: CERN NA49 Heavy Ion experiment

detail of the grid

Small, but thick scint.8 x 3.3 x 2.3 cm

Long scint. (48 or 130 cm), read out on both sides

TOF requires fast detectors (plastic scintillator, gaseous detectors), approporiate signal processing (constant fraction discrimination, corrections + continuous stability monitoring.

Time of flight


NA49 combined particle ID: TOF +

dE/dx (TPC)

T rel.

= T

/ T π

System resolution of the tile stack

L = 15 m

From γconversion

inscintillators

Time of flight


Interaction of charged particles

1

ε

ωoptical absorptive X-ray

Cherenkovradiation

ionisation transition radiation

regime:

effect:

Re ε

Im ε schematically !

• A photon in a medium has to follow the dispersion relation

222

2 022 nck

nc

knc

==−=== εε

ωλ

ππνω

εββω

θθω11

coscos ===→⋅=⋅≅nvk

kvkvrr

→ Emission of Cherenkov photons if n1≥β

Remember energy loss due to ionisation…There are other ways of energy loss !

e-

θ

khh ,ω

0,mvr

• For soft collisions + energy and momentum conservation → real photons:


Cherenkov detectors

Cherenkov radiation

Cherenkov radiation is emitted when a charged particle passes a dielectric medium with velocity

index refractive:1

nnthr =≥ ββ

lpart=βc∆t

llight=(c/n)∆tθ

wave front

1)(with1

cos ≥== λβ

θ nnnC

01

≈→= Cthr nθβ threshold

n1

arccosmax =θ ‘saturated’ angle (β=1)

Number of emitted photons per unitlength and unit wavelength interval

.with 1

sin21

12

2

2

2

22

2

222

22

constdxdE

NdEhcc

dxdNd

zn

zdxd

NdC

===∝

=

−=

νλ

λλ

θλ

απβλ

απλ

dN/dλ

λ

dN/dE

Ε

θC


Cherenkov detectors

Energy loss by Cherenkov radiation small compared to ionization (≈1%)

medium n θmax (β=1) Nph (eV-1 cm-1)air 1.000283 1.36 0.208isobutane 1.00127 2.89 0.941water 1.33 41.2 160.8quartz 1.46 46.7 196.4

( )dEEEc

LNi

i

E

EQep ∏∫= εε

αθ

2

1

)(sin2.. h

Number of detected photo electrons

EcmeVN total ∆⋅⋅= −− ε110 370

∆E = E2 - E1 is the width of the sensitive window of the photodetector (photomultiplier, photosensitive gas detector...)

Example: for a detector withand a Cherenkov angle ofone expects photo electrons

cmLeVEtotal 112.0 =⋅=∆ε

18.. =epN

o30=Cθ


Cherenkov detectors

Particle ID with Cherenkov detectors

( )222

22

11

1

11

pmn

nN

+⋅−=

−≈β

PM

particle

mirrorradiator medium

principle

Example:study of anAerogelthreshold detector for the BELLE experiment at KEK (Japan)

Goal: π/K separation

Detectors can exploit ...

• Nph(β): threshold detector• θ(β): differential and

Ring Imaging Cherenkov detectors “RICH”

Threshold Cherenkov detectors

(do not measure θC)

0 1 2 3 4 5 6

pkaon [GeV/c]

0.00.10.20.30.40.50.60.70.80.91.01.1

β ka

on ;

ligh

t yie

ld (

a.u.

)

n=1.03

n=1.02

n=1.01

βkaon

β kao

n


The Belle Detectoredited by S. MoriKEK Progress Report 2000-4.

Cherenkov detectors


.. . . .... . ..

Cherenkov detectors

Ring Imaging Cherenkov detectors (RICH)

RICH detectors determine θC by intersecting the Cherenkov cone with a photosensitive plane

→ requires large area photosensitive detectors, e.g.• wire chambers with photosensitive detector gas• PMT arrays

(J. Seguinot, T. Ypsilantis, NIM 142 (1977) 377)

+⋅=

⋅=

=p

mpnp

EnnC

221arccos1arccos1arccosβ

θ

θβ σθ

β

σ⋅= tan

Detect N photons (p.e.) →..

..

ep

ep

Nθ

θσ

σ ≈→ minimize σθ→ maximize Np.e.

n = 1.28C6F14 liquid

n = 1.0018C5F12 gas

π/K π/K/p K/p

π/h π/K/p K/p

DELPHI


Cherenkov detectors

A RICH with two radiators to cover a large momentum range. π/K/p separation 0.7 - 45 GeV/c:DELPHI and SLD

Two particles from a hadronic jet (Z-decay) in the DELPHI gas and liquid radiator + hypothesis for π and K

(W. Adam et al. NIM A 371 (1996) 240)

Principle of operation of a RICH detectors

2 radiators + 1 photodetector

C6F14 (1 cm, liquid)

C5F12 (40 cm, gas)C4F10 (50 cm, gas)

spherical mirror

PhotodetectorTMAE-based

DELPHI RICH


Cherenkov detectors

The mirror cage of the DELPHI Barrel RICH (288 parabolic mirrors)


Cherenkov detectors

“Marriage” of mirror cage and central detector partof the DELPHI Barrel RICH.


Cherenkov detectors

Performance of DELPHI RICH (barrel) in hadronic Z decays

Liquid radiator gas radiator

pK π K

p

(E. S

chyn

s, P

hD th

esis

, Wup

pert

al U

nive

rsity

199

7)

π


Cherenkov detectors (backup)

radiator (C 6F14)

quartz window (2mm)collection mesh

70 mm

cathode meshsense wires, 4 mm spacing

cadhode pads (8x8mm 2),CsI deposited

Printed Circuid Board (PCB)

2 mm2 mm

Photo detectors for RICH counters

� Gas based detectorsAdmix photosensitive agent (TEA, TMAE) to detector gasexample DELPHI:

- 54 kV

MWPC

TPC like detector (CH4 + C2H6 + TMAE)

C cone^

x-y projectionz from drift time

quartz box

example ALICE: A MWPC with CsI deposited cathode pads

Drawbacks:- slow response because of long drift times (many µs)- λ thr(TMAE) = 220 nm => only sensitive to UV light

Fast response (<100 ns), but still only sensitive to UV light


� Vacuum based detectorsPhotomultiplier tubes, Hybrid Photo Diodes.example HERMES (DESY)

3870 PMTs, Philips XP1911/UV, Ø 3/4”+ light funnels

ca. 2.5 m

3D view of one half detector

Double radiatorC4F10 gas Aerogel

Online Event Display

σθ ~ 8 mrad

Np.e. (aerogel) = 8

Np.e. (C4F10) = 12

“gas ring”

“aerogel ring”



Standoff Box 10752 PMTs in water

Magnetic Shield

Mirrors

Support tube for 12 barboxes containing 12 fused silica bars each

e- (9.0 GeV)

e+(3.1 GeV)

Bucking coil

The DIRC of BaBar(asymmetric B-factory at SLAC)Detector for Inernally Relfected Cherenkov light

e-

e+

Quartz Barbox

Standoff box

Compensating coil

Support tube (Al)

Assembly flange‘exploded’ view of the

DIRC

P. Cole et al, NIM A 343 (1999) 456I. Adam et al., SLAC-PUB-8345



Basic principle of the DIRC

Transport of C-light by internal reflection in quartz bar.Detection by PMT array. Stand off tank filled with water for ref. index matching.

u Nquartz = 1.47,u θcrit. = 43º, θc

max = 47ºu T = 99.9 % / mu R = 99.96 %u ~200-400 bouncesu loss ~10-20% = f(θdip)



DIRC: Event reconstruction is a bit more difficult !Cherenkov “ring” is decomposed into disjointed segments.

uσθ = 9.8 mrad, (7.0 mrad from ØPMT=28 mm, 5.4 mrad fromdn/dE).

uNp.e. ~ 30 uπ/K separation of 3.1 σ at p = 3 GeV/c



Transition radiation detectors


radiators) (plastic eV20 frequency

plasma

31

0

2

≈

=

∝=

pe

ep

p

meN

WW

ωε

ω

γγωα

h

h Only high energy e± will emit TR. Identification of e±

medium vacuum

Electromagnetic radiation is emitted when a charged particle traverses a medium with a discontinuous refractive index, e.g. the boundaries between vacuum and a dielectric layer.

A (too) simple picture

m Radiated energy per medium/vacuum boundary

(there is an excellent review article by B. Dolgoshein (NIM A 326 (1993) 434))

TR predicted by Ginzburg and Franck in 1946

(G. Garibian, Sov. Phys. JETP63 (1958) 1079)

A correct relativistic treatment shows that…

electron



1371

≈∝≈ αωh

WN ph

m Number of emitted photons / boundary is small

Need many transitions → build a stack of many thin

foils with gas gaps

m Emission spectrum of TR

Typical energy: γωω phh 41≈

→ photons in the keV range

• Simulated emission spectrum of a CH2 foil stack

γθ 1∝

m X-rays are emitted with a sharp maximum at small angle

→ TR stay close to track



TR Radiators:

u stacks of CH2 foils are usedu hydrocarbon foam and fiber materials Low Z material preferred to keep re-absorption small (∝Z5)

• Detector should be sensitive for 3 ≤ Eγ ≤ 30 keV.

• Mainly used: Gaseous detectors: MWPC, drift chamber, straw tubes…

• Detector gas: σphoto effect ∝ Z5

→ gas with high Z required,e.g. Xenon (Z=54)

TR X-ray detectors:

R D R D R D R D sandwich of radiator stacks and detectors→ minimize re-absorption

Intrinsic problem:detector “sees” TR and dE/dx

Pul

se h

eigh

t (1

cm

Xe)

t

dE/dx≈200 e-

TR (10 keV)≈500 e-

Discrimination by threshold


ATLAS Transition Radiation Tracker

A prototypeendcap “wheel”.

X-ray detector:straw tubes (4mm)(in total ca.400.000 !)

Xe based gas

TRT protoype performance

Pion fake rateat 90% electrondetection efficiency:

π90 = 1.58 %



Summary:

u A number of powerful methods are available to identify particles over a large momentum range.

u Depending on the available space and the environment, the identification power can vary significantly.

A very coarse plot ….

p [GeV/c]10-1 100 101 102 103 104

RICHdE/dxTOF

TR

π/K separation

e± identification

u dE/dx measurementasai/work/Particle_detector...CERN Summer Student Lectures 2002 Particle...

Documents

Transcript of u dE/dx measurementasai/work/Particle_detector...CERN Summer Student Lectures 2002 Particle...